Dan Card: AI Hasn’t Changed Cybersecurity As Much As You Think

Ravid Shwartz Ziv and Allen Roush talk with cybersecurity practitioner Dan Card about how AI changes practical attack and defence. Card discusses experiments with attacking and defending agents, automated reconnaissance, alert triage and the limits of large agent swarms. They distinguish finding a vulnerability from exploiting it in a real system, and examine social engineering, isolation, prompt injection and physical risks from autonomous systems.